Questions tagged [semi-supervised-learning]

Semi-supervised learning refers to machine learning tasks using a mix of labeled and unlabeled data. The goal is to learn a mapping from inputs to outputs, or to obtain outputs for particular unlabeled inputs. The unlabeled data is used to learn about underlying structure of the inputs, which can improve learning about the relationship between inputs and outputs. Semi-supervised learning involves elements of both supervised and unsupervised learning.

Filter by
Sorted by
Tagged with
3 votes
1 answer
91 views

Is the likelihood for Gaussian mixture models still multimodal when Y is partially observed?

In discussing Gaussian mixture models (GMMs), https://normaldeviate.wordpress.com/2012/08/04/mixture-models-the-twilight-zone-of-statistics/ highlights the issue of Multimodality of the Likelihood. ...
user avatar
  • 3,768
1 vote
0 answers
36 views

Weak supervision vs semi-supervised learning

What exactly is the difference between these two and when should they be used? Context : I have a large set of unlabelled data. I can get a number of weak labels / labeling functions for all of the ...
user avatar
1 vote
0 answers
17 views

What kind of semi-supervised learning method should be used for a low quality data set?

Consider a binary classification problem, there are $1000$ samples in the data set, of which $500$ positive and negative samples each. Positive samples have the label $1$ and negative samples have the ...
user avatar
1 vote
0 answers
9 views

About the mean-teacher algorithm, in the end we should use the student or the teacher model?

It may sound silly, but I didn't find an official approach about which final model I should choose when the mean-teacher training ends, even on the original paper. I know both will have very similar ...
user avatar
0 votes
1 answer
65 views

Semi-supervised learning: Classification vs Clustering

In the context of a semi-supervised learning problem, what's the difference between using a classification algorithm vs a clustering algorithm? Traditionally classification is supervised and ...
user avatar
  • 101
0 votes
0 answers
23 views

What is Fine Tune in "Matching Networks for One Shot Learning" paper

I was going through this paper and I understood the concept of the Meta-learning framework and using a few-shot technique. But when I tried to interpret the results in "Table 1: Results on the ...
user avatar
1 vote
0 answers
27 views

What is the best approach: Labeled training data and unlabeled test data [closed]

I'm new into the data science world and I am working on improving my knowledge so here is my problem: I want to build a binary classifier with the following constraints: I have 2 files training.csv ...
user avatar
0 votes
0 answers
11 views

Active logits length differ after augmentation in mean teacher setup for semi supervised learning

I am planning to apply mean-teacher for my problem of token classification. Since adding different noise for teacher and student is really important for the approach, i am confused about how to ...
user avatar
0 votes
0 answers
12 views

Consistency for the estimator in a mixture of objective function

Current we have two discrepancy functions $f_1(x_1,x_2,y_1,y_2)$ and $f_2(x_1,y_1)$. $f_1$ reaches minimum when $x_1=y_1$, $x_2=y_2$; $f_2$ reaches minimum when $x_1=y_1$. We consider an objective ...
user avatar
0 votes
1 answer
91 views

Computing a prior from two components in Naive Bayes

Given a model parameter $\theta$ that is composed of two distributions in a Naive Bayes classifier, how is $P(\theta)$ typically computed in practice? More specifically, from the article of Nigam et ...
user avatar
1 vote
0 answers
26 views

Improving synthetic oversampling with unlabelled data

I am working on a classification problem with a small amount of labelled data (~200 instances) and a larger sample of unlabelled data (~500 instances). To increase the size of the training data I am ...
user avatar
0 votes
0 answers
14 views

Add many unlabelled data always improve model performance for a task with a few labeled data

Currently doing some semi-supervised learning research. I am curious to this question and I think it is true since add more data help model get more information hence improve performance. Any comment ...
user avatar
  • 1
0 votes
0 answers
11 views

Why is the multiple instance learning equivalent of a problem giving much less accuracy?

I am currently working with a dataset of brain images. I have all of them labelled (0: healthy, 1: not healthy) so that I can train a fully supervised model on them. Furthermore, I have them grouped ...
user avatar
0 votes
1 answer
77 views

Binary Classification with a third 'uncertain' class label

Consider the task of classifying an image into two classes: Image shows a cat; Image shows no cat. A data set is provided for training/testing a binary classifier. However, three labels are provided ...
user avatar
  • 1
2 votes
1 answer
100 views

Semi-supervised classification objective from Kingma et al

In this 2014 paper, Kingma et al. develop different methods to do semi-supervised learning with VAEs. In one of their proposed solutions ("M2"), they approach this problem by incorporating ...
user avatar
2 votes
1 answer
22 views

Can I construct a target variable out of correlates and proxies when my training data does not have the actual target variable that I need?

I want to classify customers who at risk to churn (unsubscribe). The typical path would be to have a training set of historical data that includes observations of customers who churned so that we have ...
user avatar
  • 587
0 votes
0 answers
18 views

Semi-supervised learning from longitudinal data

I wanted to get experts' opinion on the following scenario. I am analysing longitudinal biomarker data where serial measurements were collected over time and in the last time point we know exactly who ...
user avatar
  • 611
0 votes
0 answers
24 views

Is There a General Confidence Measure for Single-View Semi-Supervised Regression?

I am experimenting with semi-supervised regression, but I cannot figure out how to compute confidence measures for the unlabeled dataset. I am reading a review paper, and it says that I am supposed to ...
user avatar
  • 112
1 vote
0 answers
129 views

What are the SOTA Visual Representation Learning architectures for binary images?

I want to learn the visual representation of binary images such as: This may later be used for the shape classification problem. I have read 2 state-of-the-art visual representation learning ...
user avatar
0 votes
0 answers
21 views

Conditional log-likelihood in Semi-supervised learning by entropy minimization

learning about the basics of semi-supervised approaches I stopped on a the paper "Semi-supervised Learning by Entropy Minimization" [1]. The Equation (1) is not so clear to derive whereas ...
user avatar
  • 101
1 vote
0 answers
22 views

Lots of unlabeled data and small set of labelled data of one class [closed]

Does anyone have suggestions for specific algorithm or implementation for labeled data of only one class and unlabeled data that can be from either classes? And I'm unsure what is the proportion of ...
user avatar
  • 11
0 votes
0 answers
25 views

Poor accuarcy score for Semi-Supervised Support Vector machine

I am using a Semi-Supervised approach for Support Vector Machine in Python for the image classification from PASCAL VOC 2007 data. I have tried with the default parameters from the libraries and also ...
user avatar
  • 259
1 vote
1 answer
36 views

Is it possible to alter binary classification models to do postive-unlabeled learning in Pyspark?

I'm learning how to use pyspark, and I'm wondering if it has any ways to implement positive-unlabeled learning? From searching this question I haven't been able to find any examples specific in spark ...
user avatar
  • 69
1 vote
0 answers
53 views

Semi-Supervised Hierarchical Topic Model

Problem statement: I'm looking to label some data with topics. These topics have a hierarchical structure (3 layers deep at maximum, but I have leaf nodes 2 layers down as well) that I have been given....
user avatar
1 vote
0 answers
17 views

Problem with a dataset not being properly labelled

I have a labelled dataset but these classes are not perfect. Some classes should be combined into one, whilst others have too few data-points for training. My main concern is the former not the latter....
user avatar
0 votes
0 answers
19 views

semi-supervised classification with a single label

I have a dataset of 1800 entries with about 40 features (some binary, some numerical). Of the 1800, 12 are known to be good for my goal; and the rest are unknown. Of the 1800 only 25-30 of the entries ...
user avatar
  • 1
1 vote
1 answer
74 views

Q-function in Q-Learning

I ran into solved old-exam question as follows: My notes tell me that option b is correct but I think option d is correct. is there any idea why (b) is correct?
user avatar
1 vote
2 answers
176 views

Semi Supervised learning vs Supervised

I am trying to understand the mathematical properties of supervised learning and semi-supervised learning. Let us consider the case for the mean $\mu$. Then the supervised learning estimator can just ...
user avatar
2 votes
1 answer
131 views

What is noise-tolerant learning?

I was reading this Development and validation of phenotype classifiers across multiple sites in the observational health data sciences and informatics network and came across the below paragraph. Can ...
user avatar
  • 1,804
2 votes
1 answer
108 views

Evaluating multiclass imbalanced problem per class

For a multiclass imbalanced problem, accuracy is not a good metric to evaluate model performance. Equally, accuracy is a global ...
user avatar
  • 231
1 vote
1 answer
505 views

Training samples with no labels: To include or not to include?

I am working on a multi-label classification problem. Each sample is capable of taking more than a single label. Sometimes samples don't have any labels associated with them. My dataset has 50% ...
user avatar
0 votes
0 answers
19 views

Latest research and explanation on how semi-supervised learning is performing better than supervised?

So in AAAI 2020 also semi-supervised learning is given the push. There are some intuitive reasoning provided by people but since the research is so fast, I wanted to know actually what is the latest ...
user avatar
2 votes
0 answers
200 views

Evaluating Semi-supervised Learning

Is there a standard procedure to evaluate a semi-supervised learning algorithm? Say if I have a set of labelled data (500 spam & 500 non-spam), and a set of 50,000 unlabelled data. Theoretically, ...
user avatar
1 vote
1 answer
24 views

Active learning with a unlabelled pool - standard references & model-based labelling of the pool?

I'm looking into active learning for a multi-class classification problem, where there is a large pool of unlabelled data. I start out with a small set of labelled data and can labelled some more of ...
user avatar
  • 22.3k
28 votes
2 answers
14k views

What's the intuition behind contrastive learning or approach?

Maybe a noobs query, but recently I have seen a surge of papers w.r.t contrastive learning (a subset of semi-supervised learning). Some of the prominent and recent research papers which I read, which ...
user avatar
1 vote
1 answer
123 views

Semi supervised learning with partially unobservable labels

As I understood the concept of semi-supervised learning is to train a classifier on the minimal available subset of correctly labeled data in order to predict the labels of a greater previously ...
user avatar
1 vote
3 answers
2k views

Can we say that RNN for time series is an example of semi-supervised learning?

I am learning neural nets, esp. focusing on RNN for my research problem. This question has nothing exactly to do with my research. With my understanding of RNN, I can think of it as an example of ...
user avatar
  • 382
1 vote
1 answer
187 views

Does it make sense to use feature selection methods to reduce dimensionality for unsupervised clustering?

If I have a dataset that is labeled with positive and negative examples, and I'd like to cluster (i.e. unsupervised) only the positive examples, does it make sense to reduce dimensionality using ...
user avatar
  • 111
1 vote
1 answer
60 views

What is it called to cluster some inputs, then classify other inputs into those clusters?

I am learning about the problem of whole-book recognition, which is tangential to optical character recognition. Some of the strategies used to identify printed characters rely on first unsupervised ...
user avatar
2 votes
1 answer
45 views

How do you train a model on a dataset that's unlabeled but we know the percentage of every class?

Say we have a data set that's pictures of apples and oranges, but we don't know which is which. However the data is organized in such a way, that for some groups of images we know how many of them are ...
user avatar
  • 123
1 vote
0 answers
35 views

Need some help understanding the factorised posterior in semi-supervised generative modelling

I am having a bit of trouble with the derivation in Kingma's semi-supervised generative modelling paper for the M2-model. The M2 model assumes a probabilistic model where the data $x$ is generated by ...
user avatar
1 vote
0 answers
40 views

Filling in Missing Data for Biological Experiment

I am trying to implement a semi-supervised learning model with biological data. In my case, I'm using features from DNA. I have a number of experiments each with many observations. Each observation ...
user avatar
1 vote
0 answers
115 views

In semi-supervies learning, is "low density separation" the same thing as "pseudo-labelling"?

I'm looking into the different methods of semi-supervised learning. In the wikipedia page, one of the methods described is called "low-density separation", where we attempt to minimize this loss ...
user avatar
1 vote
2 answers
261 views

Semi-supervised objective function VAE

In Kingma's paper on Semi-supervised learning https://arxiv.org/pdf/1406.5298.pdf, we are shown equations for the ELBO for the semisupervised case, however I am having a hard trying to derive the math ...
user avatar
1 vote
1 answer
142 views

Weak Supervision - training generative model without knowing the true label

Recently I've been reading about weak supervision. I understand most of the concept details, there's one thing that is not clear to me though. In the generative model part (creating generative model ...
user avatar
  • 769
2 votes
1 answer
597 views

Is there something more effective than ladder networks for semi-supervised learning?

The paper Semi-Supervised Learning with Ladder Networks by Rasmus is famous and interesting but a bit old now. Did researchers find any better option for semi-supervised learning ? For example, what ...
user avatar
1 vote
0 answers
311 views

What derivative to use in Gradient boosting decision tree for a semi-supervised model

I am trying to build a semi-supervised prediction model with a Gradient Boosting decision trees. The learning phase is done using the following input: $X \in \mathbb{R}^{n\times p} $ $O(X) \in \...
user avatar
1 vote
0 answers
150 views

influence of oversampling on Semi-supervised multi-label learning

I have suggested a semi-supervised approach for the hierarchical multi-label classification task. I have included the MLSMOTE oversampling technique as a pre-processing step, and then evaluate the ...
user avatar
1 vote
0 answers
19 views

What is the convergence criteria of a semi-supervised learning algorithm

I would like to know about when to stop doing semi supervision? for example, I learn a classifier from a small dataset, and then I use it to label a pool of unlabelled dataset. How long should this ...
user avatar
  • 292
0 votes
1 answer
172 views

Inference can be the goal of an unsupervised learning method or a semi-supervised learning method or even more of a reinforcement learning method?

I am new to machine learning, and I am reading a pair of machine learning books. These references talk about 2 different learning approaches: Prediction and inference, I understand the difference ...
user avatar